Exercises for the 1st CASSDA school Release 1.0 Tiago MD Pereira

Transcription

Exercises for the 1st CASSDA school Release 1.0 Tiago MD Pereira
Exercises for the 1st CASSDA school
Release 1.0
Tiago M. D. Pereira
April 25, 2015
CONTENTS
1
Introduction
1
2
Exercise questions
2.1 IRIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 CRISPEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
2
3
3
Tutorials
3.1 IRIS xfiles . . . . . . . . . . . . . . .
3.2 IRIS Dopplergrams . . . . . . . . . . .
3.3 Extracting Mg II parameters . . . . . .
3.4 CRISPEX with IRIS rasters . . . . . .
3.5 CRISPEX with spectropolarimetric data
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5
5
8
12
13
15
i
CHAPTER
ONE
INTRODUCTION
These exercises and tutorials were prepared for the 1st CASSDA School for Solar Observers in April 2015. They
cover the author’s lectures on IRIS data analysis and CRISPEX.
The questions are rather extensive and it is not expected that the attendees finish all of them during the tutorial
time. They are free to do them later, in their own time. The answers to most of these questions will not be found
in the lecture notes. The lecture notes provide only a short summary of the more extensive IRIS documentation,
which includes the instrument paper and the IRIS Technical Notes (ITNs). For many of the questions, the users
are encouraged to look in these other resources. Some of the questions will also require the user to search in the
IRIS webpage, examine some of the data files, and run some IDL commands. For these it is assumed that the user
has access to a machine with IDL and the solarsoft package installed (with the IRIS branch).
The tutorials are meant to be a guided starting point to data exploration and analysis with IRIS. They provide lists
and descriptions of a few basic tasks, and the users and encouraged to explore further and advance from them.
1
CHAPTER
TWO
EXERCISE QUESTIONS
The following list of questions is a test of your knowledge of IRIS and CRISPEX.
2.1 IRIS
2.1.1 What is the recommended IRIS data level for new users?
1. Level 0 data
2. Level 1 data
3. Level 2 data
4. Level 3 data
2.1.2 Which of the following statements is TRUE?
1. Level 2 data is level 1.5 data with cosmic rays removed
2. For SJIs there is no level 3 data
3. For spectral rasters, level 3 data cubes have a maximum of three dimensions
4. Level 1 data are flat-fielded and dark-subtracted
2.1.3 Were there limb observations on the 1st of March 2014?
2.1.4 Which of the following statements is TRUE?
1. It is possible to observe simultaneously in the FUV and NUV slit-jaws
2. The planning for IRIS observations takes place once per week
3. IRIS level 1 data is only available from the University of Oslo
4. The eclipse season of IRIS is from November to February
2.1.5 To coordinate observations with IRIS when must one let the planner know
the targets?
1. One hour before the observations
2. A week before the observations
3. By 11:00 Pacific Time the day before
4. By 00:00 UT the day before
2
Exercises for the 1st CASSDA school, Release 1.0
2.1.6 What is the approximate temperature coverage of the spectral lines observed by IRIS?
1. 10,000 K to 20,000 K
2. 5,000 K to 50,000 K
3. 4,500 K to 10,000,000 K
4. 10,000 K to 500,000 K
2.1.7 Which of the following statements is FALSE?
1. The IRIS 279.6 nm SJI is the best channel to align with the AIA coronal channels (17.1, 9.3, 21.1 nm, etc.)
2. The IRIS 140.0 nm SJI is the best channel to align with the AIA 170.0 nm channel
3. The IRIS 283.2 nm SJI is the best channel to align with the HMI continuum image
4. The WCS keywords in the IRIS file headers are obtained by cross correlation of the slit-jaw images with
AIA
2.1.8 Which of the following statements is TRUE?
1. The IRIS spectral resolution in the FUV is 53 mÅ
2. The IRIS spectral resolution in the NUV is 53 mÅ
3. The IRIS pixel size in the FUV is 53 mÅ
4. The IRIS pixel size in the NUV is 53 mÅ
2.1.9 Which of the following statements is FALSE?
1. The maximum field of view of the IRIS SJI is 175x175 squared arc sec
2. From the IRIS webpage one can only dowload level 1 data
3. The wavelength calibration of IRIS is performed by fitting the positions of known spectral lines
4. There are more than 400 Fe I lines in the NUV window of IRIS
2.1.10 Which of the following statements is FALSE?
1. In the umbra of sunspots and very strong active region plage, the Mg II k & h lines have a nearly Gaussian
shape
2. The Mg II h & k lines have an accompanying triplet of lines at 279.16, 279.87, and 279.88 nm
3. In the average quiet sun the Mg II k2r peak is stronger than the k2v peak
4. The Mg II k line is stronger than the h line and therefore it is formed in higher layers
2.2 CRISPEX
2.2.1 What data formats cannot be used with CRISPEX?
1. La Palma cube format
2. IRIS level 3 FITS files
2.2. CRISPEX
3
Exercises for the 1st CASSDA school, Release 1.0
3. Any FITS file
2.2.2 Which of the following statements is TRUE?
1. The im files contain the images, while the sp files contain the spectra
2. The sp files are a transposed version of the im files for faster reading
3. The im files contain Stokes I, while the sp files contain Stokes U, Q, V
4. The sp files must always be used, while the im files are optional
2.2.3 What does the main CRISPEX window show?
1. The intensity (or Stokes U, Q, V) at a given wavelength, animatable in time
2. The slit-jaw or filtergram intensity
3. A spectrogram of wavelength vs. time
4. Detailed spectra for the selected mouse position
2.2.4 How can one load data for multiple lines in CRISPEX?
1. Using a main file and a reference file
2. Concatenating several lines in the wavelength axis in IRIS level 3 files
3. All of the above
4. None of the above
2.2.5 Which of the following statements is FALSE?
1. CRISPEX uses the WCS keywords in the IRIS files to calculate the solar (x, y) coordinates
2. With IRIS files CRISPEX loads the first timestep into memory to calculate scaling factors
3. The y scale of the detailed spectrum window can be adjusted in the Displays tab
4. The maximum animation speed is 10 frames per second
2.2. CRISPEX
4
CHAPTER
THREE
TUTORIALS
These tutorials comprise step-by-step instructions and exercises to perform some tasks with IRIS data and
CRISPEX. As time allows, these will be done by participants in the tutorial session. It is assumed that
the participants have copied the supplied data file to their computers (if not, ask!). The file is called
data_iris_crispex.tar.bz2 and contains several datasets from IRIS and one from SST. Throughout
the tutorials it is assumed that this file is unpacked into a directory called ~/data_temp/. If you want to use a
different directory, just replace the name when necessary.
Note: Please don’t download the data for the tutorials yourself, use the supplied data to avoid any network
bottleneck.
Warning: Please make sure you have at least 10 Gb of free disk space (ideally 15 Gb) to keep all the data and
temporary files.
To start, unpack the files into a directory in your computer:
% mkdir ~/data_temp
% mv data_iris_crispex.tar.bz2 ~/data_temp
% cd ~/data_temp
% tar jxvf data_iris_crispex.tar.bz2
(...)
This will create two directories, iris and sst that have the data. You can now remove the original file
data_iris_crispex.tar.bz2.
You will need to make sure that you have a up-to-date version of solarsoft IDL with the IRIS branch. You’ll need
to have included iris in the SSW_INSTR environment variable, e.g.:
setenv SSW_INSTR ’iris hessi xrt aia eit mdi secchi sot eis’
3.1 IRIS xfiles
Go to the directory with the IRIS data, open solarsoft IDL and launch iris_xfiles:
% cd ~/data_temp/iris
% sswidl
(...)
IDL> iris_xfiles
In the middle panel, next to Search Directory: press Change.
Then navigate to the directory
~/data_temp/iris/20131226_171752_3840007146, press OK when this directory is selected. Notice that Search Pattern changed to free search.
Now press Start Search, and you should see the list of files appear. Feel free to check the slit-jaw movies
and then double click on the raster file. What can you say about the line list?
5
Exercises for the 1st CASSDA school, Release 1.0
Select the Mg II k 2796 line and under Line fit select Profile Moments. On the Moments Prep Tool window
adjust the reference wavelength so that it matches the k3 core, and set the line start and stop so that it is about 5
pixels wide around k3. Set the continuum to be about 5 pixels wide on the right side of the plot, in a region with
no absorption lines. Press Finished. What do you see?
The result for intensity should look like this:
Now select the Si IV 1403 line instead, and choose again Profile moments. The result for intensity should look
like this (set log(HistoOpt Value) to -1.65):
Note: The single or double Gauss fit options in Line fit are not currently working.
Go back to the IRIS_Xcontrol window of the raster file. Select the Si IV 1403 and Mg II k 2796 lines and press
Generate level3 files. In the next dialog select only add “20131226_171752_3840007146/” to save directory, so
that the level 3 file is saved in the existing directory. We will use this level 3 file later with CRISPEX. Note that
no sp file was created, as these observations comprise only a single 400-step raster.
Go back to the main window of iris_xfiles.
Search in the directory
~/iris_data/level2/2013/09/14/20130914_230510_4004257647.
Feel free to explore
this data set. Then select the C II 1336 and Mg II k 2796 lines and press Generate level3 files. Then check copy
reference SJI file and press OK. We will use these level 3 files next with CRISPEX. Note that now two files were
created because this dataset has a temporal component.
3.1. IRIS xfiles
6
Exercises for the 1st CASSDA school, Release 1.0
3.1. IRIS xfiles
7
Exercises for the 1st CASSDA school, Release 1.0
3.1.1 CRISPEX
In the same directory as before (and perhaps same session) open CRISPEX and load the level3 file from the 26th
of December:
IDL> crispex, ’iris_l3_20131226_171752_3840007146_t000_SiIV1403_MgIIk2796_im.fits’
It will take a while to start as CRISPEX loads everything it needs into memory. If everything went well you can
start exploring the data with CRISPEX!
Now load CRISPEX with the level 3 files for the 14th of September 2013 and use the level 2 slit-jaw image as
reference:
IDL> crispex, ’iris_l3_20130914_230510_4004257647_t000_all_im.fits’, $
’iris_l3_20130914_230510_4004257647_t000_all_sp.fits’, $
sjicube=’iris_l2_20130914_230510_4004257647_SJI_1400_t000.fits’
After the load a warning message will pop up saying that NX=1 leads to an extreme aspect ratio. Press OK.
Explore this set. As it is a slit and stare, one can only look at the spectra long the slit, and the spectrogram plot
now displays the time component. When the time series is animated, the current time is shown as a horizontal
grey line moving on the spectrogram plot.
3.2 IRIS Dopplergrams
In this tutorial we will look at a very large dense raster in the south pole. The data set directory should be:
IDL> data_dir = ’~/data_temp/iris/20131010_100202_3820259146’
Feel free to examine these data in iris_xfiles. This very large dense raster took about one hour to complete the 400 scans, which means that the orbital velocity and thermal drifts were changed during the one hour
observations. This means that any precise wavelength calibration will need to correct for those shifts.
First lets load the data using the IDL object interface:
IDL> filename = ’iris_l2_20131010_100202_3820259146_raster_t000_r00000.fits’
IDL> filename = data_dir + ’/’ + filename
IDL> d = iris_obj(filename)
Let us see the lines that are saved in this raster:
IDL> d->show_lines
Spectral regions(windows)
0
1335.71
C II 1336
1
1349.43
Fe XII 1349
2
1355.60
O I 1356
3
1393.78
Si IV 1394
4
1402.77
Si IV 1403
5
2832.76
2832
6
2814.50
2814
7
2796.20
Mg II k 2796
Let us load the Mg II k line into memory:
IDL> wave = d->getlam(7)
IDL> data = d->getvar(7, /load)
We can see how the the spatially averaged spectrum looks like:
IDL> mspec = total(total(data, 2), 2)
IDL> plot, wave, mspec
IDL> plot, wave, mspec, xrange=[2794, 2799], /xst
3.2. IRIS Dopplergrams
8
Exercises for the 1st CASSDA school, Release 1.0
To better understand the orbital velocity problem let us look at how the line intensity varies for a strong Mn I at
around 280.2 nm, in between the Mg II k and h lines. For this dataset, the line core of this line falls around index
350. To plot it in the correct orientation we will make use of IDL’s rotate, and the procedure pih (available in
the IRIS tree of solarsoft) to make the plot:
IDL> pih, rotate(reform(data[350, *, *]), 1), min=0, max=30, scale=[0.35, 0.1667]
The result should look like this:
Figure 3.1: Intensity at Mn I 280.2 nm line when orbital velocity and thermal drifts are not accounted for.
You can see that the left side of the figure is brighter, and indication that its intensities are not taken at the same
position in the line because of wavelength shifts.
To calculate the wavelength shifts from the orbital velocity and thermal drifts we do the following:
IDL> iris_orbitvar_corr_l2, filename, corr_fuv, corr_nuv
IDL> loadct, 0
This saves the shifts (in Ångström) into the variables corr_fuv and corr_nuv (respectively, for FUV and
NUV spectra). The loadct call is to change to the default colormap. To look at intensities at any given scan we
3.2. IRIS Dopplergrams
9
Exercises for the 1st CASSDA school, Release 1.0
only need to subtract this shift from the wavelength scale, but to look at the whole image at a given wavelength we
must interpolate the original data to take this shift into account. Here is a way to do it (note that array dimensions
apply to this specific set only!):
IDL> new_data = fltarr(535, 1092, 400, /n)
IDL> .r
for i=0, 399 do begin
for j=0, 1091 do begin
new_data[*, j, i] = interpol(data[*, j, i], wave - corr_nuv[i], wave)
endfor
endfor
end
(This double loop may take a while to complete.)
Once you have the calibrated data, we can compare again how it looks at the Ni I line wavelength:
IDL> pih, rotate(reform(new_data[350, *, *]), 1), min=0, max=30, scale=[0.35, 0.1667]
And now we can see that the intensity map is uniform along the solar disk:
Figure 3.2: Intensity at Mn I 280.2 nm line when orbital velocity and thermal drifts are accounted for.
3.2. IRIS Dopplergrams
10
Exercises for the 1st CASSDA school, Release 1.0
We can use this calibrated data for example to calculate dopplergrams. A dopplergram is the difference between
the intensities at two wavelength positions at the same (and opposite) distance from the line core. For example,
at +/- 50 km/s from the Mg II h3 core. To do this, let us first calculate a velocity scale for the h line and find the
indices of the -50 and +50 km/s velocity positions:
IDL>
IDL>
IDL>
IDL>
IDL>
h_centre = 2796.41
vel = (h_centre - wave) * 3e5 / h_centre
; find index of -50 and 50 km/s
tmp = min(abs(vel - 50), i50p)
tmp = min(abs(vel + 50), i50m)
Now get the dopplergram and plot it:
IDL> doppgr = rotate(reform(new_data[i50m, *, *] - new_data[i50p, *, *]), 1)
IDL> pih, doppgr, min=-30, max=30, scale=[0.35, 0.1667]
Figure 3.3: Dopplegram for Mg II h at +/- 50 km/s.
3.2. IRIS Dopplergrams
11
Exercises for the 1st CASSDA school, Release 1.0
3.3 Extracting Mg II parameters
Here we will work with a few examples to make use of the iris_get_mg_features_lev2, which finds the
position of the Mg II h&k features.
3.3.1 Active region plage
The first data set is a very large dense raster of active region plage. The files can be loaded like this:
IDL> data_dir = ’~/data_tmp/iris/20131226_171752_3840007146’
IDL> filename = ’iris_l2_20131226_171752_3840007146_raster_t000_r00000.fits’
IDL> filename = data_dir + ’/’ + filename
We can calculate the properties of the Mg II k line in the following manner:
IDL> iris_get_mg_features_lev2, filename, 3, [-40, 40], lc, rp, bp, /onlyk
The output was saved in the arrays lc (line centre), rp (red peak), and bp (blue peak). To save time we calculated
only for the k line. We can then visualise both the derived velocities and intensities. For the intensities:
IDL> pih, rotate(reform(lc[0, 1, *, *]), 1), min=0, max=500,
IDL> pih, rotate(reform(bp[0, 1, *, *]), 1), min=0, max=750,
IDL> pih, rotate(reform(rp[0, 1, *, *]), 1), min=0, max=750,
scale=[0.35, 0.1667]
scale=[0.35, 0.1667]
scale=[0.35, 0.1667]
and for the velocities:
IDL> pih, rotate(reform(lc[0, 0, *, *]), 1), min=-15, max=15, scale=[0.35, 0.1667]
IDL> pih, rotate(reform(rp[0, 0, *, *]), 1), min=0, max=30, scale=[0.35, 0.1667]
IDL> pih, rotate(reform(bp[0, 0, *, *]), 1), min=-30, max=0, scale=[0.35, 0.1667]
3.3.2 Sit and stare in plage
Now we are not going to use a dense raster, but a sit and stare instead. In this program the slit does not move in
space, so the level 2 spectrograms are given as function of space and time. Let us load the set from the 14th of
September 2013:
IDL>
IDL>
IDL>
IDL>
IDL>
IDL>
data_dir = ’~/data_temp/iris/20130914_230510_4004257647’
filename = ’iris_l2_20130914_230510_4004257647_raster_t000_r00000.fits’
filename = data_dir + ’/’ + filename
d = iris_obj(filename)
wave = d->getlam(7)
data = d->getvar(7, /load)
Let us look at the average spectrum:
IDL> mspec = total(total(data, 2), 2)
IDL> plot, wave, mspec, xrange=[2794, 2799], /xst
Upon closer inspection, we see that the k3 line centre is not at its expected wavelength, but rather shifted by about
0.129 Å to the red (~14 km/s). This occurred probably because of an error in the automatic wavelength calibration.
To calculate the Mg II k features this will cause a lot of problems, as the code assumes a good wavelength
calibration. However, we can supply a wave_comp keyword to iris_get_mg_features_lev2 with the
amount of shift:
IDL> iris_get_mg_features_lev2, filename, 7, [-40, 40], lc, rp, bp, $
wave_comp=-0.129, /onlyk
Note: The wave_comp keyword is not available in earlier versions of iris_get_mg_features_lev2, so
be sure to have an updated version.
3.3. Extracting Mg II parameters
12
Exercises for the 1st CASSDA school, Release 1.0
Figure 3.4: Mean spectrum for plage sit and stare.
We can now look again at velocities and intensities:
IDL>
IDL>
IDL>
IDL>
IDL>
pih,
pih,
pih,
pih,
pih,
rotate(reform(lc[0,
rotate(reform(bp[0,
rotate(reform(bp[0,
rotate(reform(rp[0,
rotate(reform(lc[0,
0,
0,
1,
1,
1,
*,
*,
*,
*,
*,
*]),
*]),
*]),
*]),
*]),
1),
1),
1),
1),
1),
min=-15, max=15
min=-35, max=10
min=0, max=700
min=0, max=700
min=0, max=500
The last plot (line centre intensity) looks like:
What are the typical lifetimes of dynamic fibrils (e.g. from the parabolic motions seen in the k3 intensity image)?
3.4 CRISPEX with IRIS rasters
3.4.1 Active region 400-step raster
Let us go back to the directory of the first dataset we worked with, and run CRISPEX on the level 3 file:
IDL> cd, ’~/data_temp/iris/20131226_171752_3840007146’
IDL> crispex, ’iris_l3_20131226_171752_3840007146_t000_SiIV1403_MgIIk2796_im.fits’
Explore the dataset with CRISPEX, and go through the following tasks/questions:
• Adjust the scaling of the spectral plot so that the lines are visible (Displays tab, lower/upper y-values, and
also multipliers in Scaling tab under Detailed spectrum)
• Look at the main image in the cores of the Mg II k and Si IV lines. Adjust scaling for Si IV 1403 so that it
becomes visible (change ‘Histogram optimisation to 0.001 and/or set gamma lower than 1)
• Blink the image between the spectral positions of the cores of the Si IV and Mg II k lines (use animation
speed of about 2 frames/s)
• Can you find a large dot where Si IV is greatly enhanced but Mg II is not too unusual? What are its solar (x,
y) coordinates?
• Is there a sunspot or a pore in this image? How do you find out?
3.4. CRISPEX with IRIS rasters
13
Exercises for the 1st CASSDA school, Release 1.0
Figure 3.5: Intensity from k3 as a function of position in the slit and time.
3.4. CRISPEX with IRIS rasters
14
Exercises for the 1st CASSDA school, Release 1.0
3.4.2 Active region sit and stare
Now let us look at a different type of IRIS observation, where the slit is fixed in space (sit and stare). We have
not yet produced the level 3 files for CRISPEX for this dataset, so let us change directory and do that from the
command line:
IDL>
IDL>
IDL>
IDL>
cd, ’~/data_temp/iris/20130914_230510_4004257647’
f = iris_files(’*raster*’)
; make level 3 files with only Mg II (index 7)
iris_make_fits_level3, f[0], [7], /sp
Let us now run CRISPEX using both im and sp files, and also use a slit-jaw image:
IDL> l3files = iris_files(’iris_l3*’)
IDL> sji = iris_files(’*SJI*’)
IDL> crispex, l3files[0], l3files[1], sjicube=sji[0]
Explore the dataset with CRISPEX, and go through the following tasks/questions:
• Adjust the scaling of the slit-jaw image (Scaling tab, select Slit-jaw image, change Histogram optimisation
and/or gamma)
• Start running the time sequence. At what time does this observation finish?
• Show only the Mg II k line (Spectral tab, narrow down Doppler minimum/maximum values)
• Lock the mouse at a given position, see how the temporal evolution goes through the Spectral T-slice window
• Can you find a location with shock signatures in Mg II? (e.g. y=193)
3.4.3 Penumbral 16-step raster
In the directory ~/data_temp/iris/20140630_205235_3820255482/ there is an IRIS dataset from a
16-step raster, with 100 repeats. In CRISPEX the visualisation of this dataset has both a time domain and a main
image with 16 columns. Feel free to explore this dataset, first by creating the level 3 files and loading them in
CRISPEX. When used with a sjicube option, one can see the 16 positions superimposed in the slit-jaw image.
3.5 CRISPEX with spectropolarimetric data
Here we will use CRISPEX with spectropolarimetric data with the full Stokes (I, Q, U, V) intensity vector. The
sample dataset was observed with the Swedish Solar Telescope (SST), and includes the Fe I 630.1 nm line. Start
by going to the data directory:
IDL> cd, ’~/data_temp/sst/’
You will find two .icube files and an .idlsave file. These files are the older (native) format of CRISPEX,
and not the same as the IRIS level 3 files. They can be loaded in the same fashion, with a different option for the
spectfile (which contains the wavelength):
IDL> crispex, ’crispex.6302.rebinned.fullstokes.icube’, $
’crispex.6302.rebinned.fullstokes_sp.icube’, $
spectfile=’wavelength.idlsave’
Explore the dataset with CRISPEX, and go through the following tasks/questions:
• Display and animate the intensity panel, then do the same for the Stokes parameters Q, U, and V
• Adjust the scaling in of the detailed spectra for both Stokes I and Stokes V
• Based on the red wing Stokes V, where are the negative and positive polarities in the sunspot?
• Where is the horizontal polartisation larger than the vertical polarisation?
• Can you find any umbral dots? (Hint: use a low gamma on the intensity images)
3.5. CRISPEX with spectropolarimetric data
15